Multimedia Information Security Expectations

 

By Roy D. Follendore III

Copyright (c) 1992 Roy Follendore III

 

Multimedia is the principal presentation and distribution technology emerging from data network and distributed processing communications.   Multimedia is also a rational progression from desktop publishing technologies.  With the advent of open system design in software and hardware, multimedia is here and is changing the expectations users have of computer information.   Because of the widespread adoption of all types of computer information into "multimedia productions,"  the security of multimedia is a hot topic.  Until recently, security experts assumed that traditional INFOSEC models, involving system high or end to end encryption, could be directly adapted to all impending multimedia environments.  The underlying security philosophies of traditional INFOSEC model remain similar to the models presented for paper cable traffic available since the late 1940's.  Localized policies regulate what can and can't be communicated.  What has changed is the basic nature of  the information to be communicated, transported, stored and retrieved.  In 1988, the Congressional Board of the 100th Congress wrote a paper entitled " Overview of Science, Technology, and the Constitution" (OTA-BP-CIT-43)  In this paper they said,  "Our way of life in the U.S has, to a large extent, been shaped by changes in science and technology.  And as we enter our third century of constitutional history, these changes are likely to continue with increased importance."   Widespread application of multimedia presents an acute problem for both outdated INFOSEC security models and outdated expectations of security in society.[1]  Any future INFOSEC solution must interact successfully with the essential qualities of multimedia, qualities that involve enhanced user performance.

Multimedia represents an aggregate of  "objects", consisting of audio, visual and textual information.  Conceptually, each "information object" has many flavors and durations.  Multimedia has an underlying objective of the transmission of more than just the symbolic information of text or literal and symbolic qualities of images and graphics.  Operationally, multimedia represents a production of values that is designed be transmitted and reproduced in concert with the cognitive nature of receiver.  Because multimedia is an engagement of interactions between the author and the receiver,  it is most complex.  The complexity of multimedia is much more than the complexity of the technologies.  Like the nature of fine art and science, for which it shall serve, it is the nature of multimedia that the whole of the qualities are greater than the sum of the parts.[2]  This is due in part because multimedia is a receiver oriented form of communication, whose effects are often more related to interactive video production than text processing.  Information in the form of standard text does not have the same human impact as well designed multimedia.  Any restriction or timing delay of the elements of multimedia can change that impact on the receiver and can significantly change the message of the author.  The question is: Can a generic multimedia INFOSEC model be developed to insure security without affecting performance? The answer is simply that it must.  Multimedia INFOSEC solutions that do not perform well enough can either be ignored or, when enforced, present a crippling overhead to operational resources.

INFOSEC models that have worked in favor of  National Security in the past are inadequate for the complexity of multimedia and our distributed processing society.  The purpose of this paper is to discuss fundamental technical implications of multimedia security and present an strong and dynamic alternative INFOSEC data separation model that can operate within future multimedia systems.  

Traditional INFOSEC models provide security for systems by using a arrangement of unyielding standards and policies that often limit the potential software applications that can be used.  Today's more advanced Commercial Off The Shelf (COTS) multimedia software  take advantage of many kinds of simultaneously input and output formats.  It is certain that these systems will become more complex and will be designed take advantage of the user's every physical sense, in order to better communicate and interact faster and more effectively.  If INFOSEC models remain fixed, as multimedia software becomes more common and more sophisticated, the INFOSEC model will, through operational necessity, be forced to relax rigid policies in favor of user performance.  If so, there could be two potential destructive  side effects.  First, rigid INFOSEC security policies could be maintained in spite of performance.   Second,  with no viable options at hand,  existing INFOSEC security policies risk being forced to relax and reduced or simply wavered for the convenience of operational production. 

 

Even without the complexity of multimedia, the United States national security policy has been confronted with a gulf of problems associated with realizing empirical validity across different organizational  INFOSEC models.[RDF1]   Cross organizational INFOSEC problems can be traced to a fundamental conflict of philosophies associated with inappropriate integration of open and closed systematic approaches to communication and information security along with  illogical data classification policies. [3]   To more clearly understand "why" these conflicts occur it is important to maintain a high view of the problems.  The assessment of information security must be  presented and  considered from open and natural organizational communications and Information Science perspectives.[4][RDF2]    From this level, information security problems most often occur because there are critical philosophical gaps in the rational for embedding INFOSEC products across open and closed communities of networks and systems doing similar work.  Traditional INFOSEC approaches often involve expensive integration of security products that protect explicit links in the communications paths between organizations .  As new technology comes and goes, new and less secure data paths arise as a byproduct of unplanned organizational engineering requirements and normal organizational change.    Because of this, many of today's INFOSEC models began by generalizations of requirements but operate by an unending series of costly "security bandaids."  Tracking "what is, as opposed to what should be" and "what is expected" becomes the all time consuming effort of an army of INFOSEC "experts."  The result under existing INFOSEC models is that as information systems become more complex so too has the model.  The more complex the INFOSEC model the more INFOSEC "experts" are required to manage the complexity.  This adds up to enormous cost to the organization in the form of overhead, loss of productivity and unknown expense in missed human to human communication.  All of this occurs because the appropriate paths for communication were not available and/or were not secure.  Inappropriate INFOSEC multimedia models could amplify this problem and, as we shall see, significantly distort the message in much the same manner as the product of poor editing of censors preparing film for television.  Multimedia will drive technologies to improving INFOSEC solutions and asking fundamental questions concerning what we know about this problem.  There are two questions that must be considered essential to providing a multimedia INFOSEC solution.  (1.) Why is  the present INFOSEC lacking?  (2.) How can a solution be best approached?

 

All INFOSEC models must formally and fundamentally deal with risk.  These risks include, but are not limited to those risks associated with the formation and maintenance of axiomatic security policies.   Present INFOSEC  policy suffers from the inability to maintain axiomatic  policies with rapidly changing technologies and shifting user needs.  Security becomes problematic and difficult for organizations to adhere when meeting operational needs.  INFOSEC can too quickly become a barrier to organizational production demand.  Pitted against the survival of organizational production, INFOSEC wavers are then either formally or informally authorized, creating uncharted craters in the INFOSEC model.    Even under ideal circumstances, barriers to information access are a continual and difficult problem and cause dangerous gaps of information provided to analysts and decision makers.[5]  

 

From the organizational view, a typical INFOSEC  model simply describes many virtues, rules and requirements of information security without directly bridging them to issues of organizational performance.  It mostly consists of do this and don't do that.  This causes at least two often conflicting views for designing information security measures.  For users and managers who are interested primarily in results, traditional information security  measures are viewed as policies that tend to degrade both individual and organizational performances.   Viewed from a the security management perspective, INFOSEC protects the opportunity of the individuals to perform through the protection of critical information.  Once an organizational INFOSEC model is designed, risk analysis to determine the continuing viability of INFOSEC is rarely performed.  The machinery of security remains in place often as artifacts for which organizations have learned to budget future resources.  The cost of doing reoccurring information systems security analysis as well as the uncertainty of occurring risks may prevail year after year.  Why not leave the existing security measures in place and integrate other measures as required?  The answer is that the organizational requirement for quality  information that is "just in time" parallels the production models for modern manufacturing.   Technologies are rapidly evolving critical definitions and applications of information.[6]  New technologies will generate new questions that shift the fundamental nature of security to quality performance issues.  Industries are openly operating in environments of ever higher competitive markets while Government programs and organizations are engaged in a dance of shifting competition for funding. 

 

The premium preoccupation of security experts has been  that security was is equated to organizational survival.  Because of the argument that organizational survival is "essential," it was easy to persuade managers of the need for "absolute" security solutions.  Security products often provide "absolute" security by limiting user performance.  Very little work has been done by security designers to intentionally improve existing organizational performance and communication quality using modern security technologies.  The traditional direction of security engineers has stressed damage control over performance improvement.  Can quality INFOSEC  models be designed and  developed not only to protect but also to enhance individual and organizational performance in an age of  computer mediated audio, video, hypertext and graphics?  Can INFOSEC be integrated so intimately with improved performance that security is virtual to the operation of the system?  The short answer is that for multimedia to be productive and successful as communications technologies in secure environments, it must!   Quality INFOSEC solutions should be a virtual byproduct of increased organizational system performance.  Modern programming techniques and strategies can reverse many of the differences and much of the cost by implementing  a fundamentally integrated and automated "Virtual INFOSEC Systems Architecture."  Virtual INFOSEC represents the expectation that the burden of future INFOSEC policy will not be born by organizational members.  INFOSEC will be hidden from the user,  encapsulated in the systems that users operate.  At the heart of  Virtual INFOSEC there is no requirement for users to be INFOSEC system aware.  The systems that users operate would take care of the INFOSEC requirement.  Taking maximum advantage of technology,  Virtual INFOSEC policies will provide empirical solutions to data categorizing and aggression issues as well as data, information, and knowledge separation.[7] 

 

To  provide a consistent philosophical baseline for multimedia security we need to reconsider the rock bottom rational, the natural foundations of INFOSEC.  We must rediscover and realign what we know of modern Information Science philosophies and in the process update our conceptions of INFOSEC.   In doing this, we can view INFOSEC free of much of the technical and political bias that often cloud policy making judgments.  There is a swarm of views concerning what makes good INFOSEC.  They range from problematic observations to forecasts.  The observation that the final objective of multimedia is direct communication of human minds through the full use of human senses is enough to set off a storm of controversy.  Statements that multimedia's final potential is the process of computer mediated communication using many human physical and cognitive senses to convey meaning and understanding (knowledge) might also be just as controversial to others.  Multimedia is a general term meaning many things to different people.   It can also be a general solution for many specific problems.  Like all general solutions multimedia invokes opportunistic responses and also resistance due to personal investment.  Multimedia is a term that implies the adaptation of technologies to presentations and interactions that are appropriate to desired physiology's and cognitive products.  Multimedia can therefore be considered akin to both psychology and physiology  and the "natural human to human" communication process.  What happens to idea of security if in multimedia technologies if we agree that multimedia is leading to a fuller and more natural way of communication using modern technologies?  

If INFOSEC can be thought of as a natural process, what can we learn from natural models? 

Signals are used to communicate information and communication of information by signaling is not unique to man.    It is therefore follows that many of the applied information security issues are not unique to humans.   Animals have been communicating, intercepting and mimicking each other naturally for millions of years.  These organisms represent communications and INFOSEC systems that interact and compete.  Certain species of fireflies, for instance, mimic the mating signals of other species, to eat them when they arrive.[RDF3]   

Information science begins as a natural science.  INFOSEC can be considered a natural part of information science.[8]   From this perspective, there is a wealth  of important information to be learned and many immediate advantages for understanding. There are many fundamental questions that can be posed for meeting our objective of rationally specifying the INFOSEC data separation model that can enhance future multimedia systems and organizational communications.

This World has always been a very dangerous place for both man and beast.  Was communication originally created to sustain organizations of interacting interdependent life?  We know that survival of an individual's genetic memory is not always enough to promote survival of a distinct species.  From the point where the first two single cells, by accident, or by act of God,  found each other, clung to each other and then thrived, there has existed a need to communicate between organisms.  For microorganisms, communication of individual existence and common needs were certainly the first requirement for cooperation and interaction between independent cells. 

Millions of years of cyclic evolutionary successes by independent cells have produced the elaborate multicelled organisms we conceptualize as independent animals.   Modern man may take for granted the fact that very dissimilar internal cells of animals communicate through biological and genetic messages to cooperate for the good of the body.  Perhaps it is the nature of civilized man to forget that the philosophies of our modern computer technologies are founded on early perceptions of these discoveries, many within our lifetime.   It is more than blind oversight to consider the act of communication between human beings within organizations and then ignore the opportunity for fundamental first principal observations found in natural organizational communications models.  If organizations of people communicating independently and in groups across boundaries securely is an INFOSEC problem, then there is much that can be learned and directly applied through observation of natural models whose successful design must be based on the successfully secure solutions of private and public transfer of information between competitive and cooperative species.  This is a point that is not within the scope of this paper and is best to be resolved in future basic research.  At the same time, the issue of natural INFOSEC is brought forward for two reasons.  First, in many ways much of modern hardware and software science technologies are designed and operate in cellular progression. Second, modern research over the last three decades have presented more appropriate models of human organizations than the traditional mechanical view of dynamics often referenced in "wiring diagrams" of authority.  These models have been well researched and are often based on Natural Systems theories found in the Natural Sciences. Converging philosophies of the physical, logical and natural sciences have driven home unique opportunities for the true integration of  INFOSEC into both the intimate functional model of the organization and the enabling technological transport.

Humans have inherited requirements for both  inter and intra organizational communication that arise from the same source of requirements as fireflies, or micro cells in our environment.  The message remains much the same, "This World is still a very dangerous place and survival of single organisms are not enough to promote the survival of a society."

 

Improved Communication Improves Organizational Sensitivity To The Environment

Security in any natural system is never as simple as it may first seem .   This is due in part to the notion that survival may have many incongruent qualities associated with success.   It is rational to assume that the animal brain evolved from the fact that simple automatic reflex actions associated with light and sound were not enough to assure success of total cell group survival.  For animals, more information could be acquired from light and audio stimulation, information that presents opportunistic solutions, and information that is secure of natural or induced external error.   Like most animals, individual human beings balance survival both as a society of communicating interdependent cells and associated interdependent beings forming communicating groups. 

Like animals, people use their brain to infer the presence of danger from all possible senses.   To survive, brains must be capable of taking a maximum of appropriate sensory activations and quickly associate the activations with the presence of environmental danger.  There is a physical limitation to the speed and sensitivity at which this process can take place, but there is also a system that improves the correctness, and therefore the survival security of the activations.   Both animals and humans naturally group together as organized supersystems to improve the potential of such sensory activations, to communicate,  and through interaction,  amplify and often re transmit (maintain) the urgency of the danger.

Organizational supersystem structures impart a high degree of sensitivity and complexity to the communications process.  An ant colony is often presented as a "body" for instance.  Organized complexity has both an overhead and a reward for all groups.   Errors associated with organized transmission and recognition communicated danger signals must be reduced for the good of the body.  The use of multiple senses, such as those found in the audio, visual and tactile senses,  reduce the prospect of imperfect environment noise by redistributing the potential of error across different dimensions of time and space.  Moreover, multiple senses act to authenticate the communicator's intentions and state of mind.    Like our ancient ancestors, modern organized people still depend on groups to amplify their senses in detection of danger and performance of critical work.  We reduce error and authenticate through multiple senses.  We build our organizations to reach common goals through communications in very similar ways as our ancestors. 

 

Nonverbal organizational action is a form of communication that greatly improves group reaction time, allowing members critical reaction time to take appropriate independent action and save itself.       Schools of fish, flocks of birds and herds of grazing animals all use organizational action as a basis of communicating danger.  For natural nonverbal organizational action to act as a form of communication  requires learning in the presence of a group in a dangerous environment.  We may consider hand gestures and movement of face muscle as nonverbal forms of communication.  Nonverbal communication is not necessarily primitive, in fact it is often very sophisticated.  Consider that many social scientists believe that over 80 percent of information transferred between people working in contact with each other is presented and received as  nonverbal information.  It is speech or text that may quickly represent procedure and concept but it seems that it is often that nonverbal glance that can often prioritize the importance or lend credence to the overall message.  The multisensual presentation of information is essential to both the convergence of the cognitive state of the communicator and to the authentication of the message.

 Did people have incentives to learn to communicate through symbols rather than just nonverbal action?   The advantages of group security humans enjoy today hinge on the fact that the actual presence of danger is no longer absolutely necessary.  Individual people can be informed of the indications of danger and understand that danger without ever having experienced it.  The communication of information through symbolic information provides groups of humans with more individuals capable of predicting more than what they personally witness. [9]   Accepted symbols of the community have come to represent or replace many of the direct actions of the group.  It appears that most truly fundamental communications are associated in some way with survival information to the individual as a part of an organization.  Security information is still the principal obligation an individual maintains as a member of  any organization.

A lot of communication is necessary in organizations that insist on ordered information in propriety.  Clearly the largest and most important commodity of our modern society is information. Without good information, humans can't make good decisions.  This observation doesn't mean that good results can't come from bad decisions, or that bad results don't come from the best of decisions.   Humans don't always accept or act on good or bad information.  The maintenance of the proper supply of good and truthful information allows individuals more opportunities to improve upon independent decision  making and thus their independent survival prospects.  Information systems that attempt to impose external INFOSEC solutions on individuals, where immediate organizational survival directly conflicts with individual survival, are fundamentally flawed. 

The process of information sharing is critical for the very existence of  organized people.  Simplistic notions that organizational decisions are directly made and actions are directly taken does not consider the requirement of communication for facilitating planning, decision making and ultimate actions.  Organizations are designed more to facilitate the communication process than for any other reason.  The ways in which societies are ordered affect the production and sharing of information.  Trustworthy information provides mankind with the quality of life associated with informed choices.   Those activities surrounding the sharing of information are often the predominate processes of organizations. 

History effectively demonstrates that organizations and societies as well as individual humans and animals compete for different qualities necessary for survival.  Societies of our time produce and share immense quantities of information.    Organizational communication can be seen as a series of networked objects with tangled dynamic pathways that stream together for common purpose.   When viewed from these higher perspectives it is the organizational communications pathways that represent the true formal organizational communication boundaries.

From a technological perspective, data communication begins as binary elements representing either an "on" or "off," "yes" or "no."  Enough "yes" or "no" data elements and  it is the complex association of "yes" and "no" that can be considered as information.   The act of association requires the act of remembering previous data elements.  Loss of one or more data elements changes the context of the information and therefore the meaning.  It is the association of these elements together and the individual's memory of data elements in data groups that give meaning not provided by any single data element alone. 

At the same time, most data must be lost for information to be understood.  This is because it simply isn't true  that data alone represents information.   Most data in the World is simply representing noise from an individual's perspective.  The choice of what is noise and should be lost and what isn't and should be maintained is vital to the information in context.

Data is useful only if exploitable by the assumption that there are detectable relationships to other data along with further refinements and evaluations that can be transformed into new information.  Data must be presented in context with other data to make sense as a body of information.[10]  For data to become information, other relative data must preexist and must therefore be accountable.  It is therefore this associated matrix of data that becomes information. 

Using similar reasoning, it is an associated matrix of information that becomes knowledge.  Knowledge is thus built of information, which in turn is built upon data as differentiated from noise.  The lowest of organisms might function well with data as their highest order of communications.  Larger and more complex animals must improve their survival performance by responding to information. 

Humans attempt to use knowledge as the basis of formulating  communications and actions.  Organizations in turn rely on intricate inter and intra personal relationships to organize the rational transfer of information and act on the formulated knowledge.   Along with the increase in complexities of organizations have come increased complexities in the communications process itself. 

Denial of internal organizational information can be the result of guileless organizational or individual error, or the intentional falsification or expunging of data and/or information.  People can't freely act for or against things they don't know of, but people can and often do act on information that is incomplete and/or incorrect.

The Protection of the integrity and distribution of information through organizational communication channels is vital to the transformation of knowledge into organizational action.  All information, formal and informal,  operates to reduce uncertainty and introduce bias for organizational members to act or not to act.    

The intentional falsification or expunging of organizational information is the modification of organizational knowledge.  Falsification changes information and therefore knowledge during knowledge transfer.    Expunging information also modifies knowledge as transferred.  Thus the same organizational consequences are found both in altered information and completely fabricated information.  

Expunged information may lead to gaps in information that are detectable.  Altered information carries an additional distinction of often being more difficult to initially detect  simply because the information may be partially true.  The parts that are true could have originated within legitimate organizational channels.   Altered information carefully inserted can be much more difficult to correct in organizations because it can be tailored to support existing individual and organizational bias for or against change.

The methods and qualities associated with the classification of information are intimately tied to the organizational information flow process.  Information is usually qualified relative to it's known function.  Generally this qualification is associated to the accomplishment of a particular organizational mission.  In many organizations, however, the idea of information classification is pinned to the organizational structure.  Analysts in intelligence communities, for instance, are often required to place classifications on information simply because they wrote a document.  Their position in context to the material becomes the overriding criteria for the classification process.  As we can see, local policy can irrationalize the process of information classification where information sensitivity and organizational structure are also at stake.

It is often said in such organizations that where everything is classified, nothing is classified.

At issue here is not just the logic and rationality of the data classification process, but what is more important for INFOSEC, the central question of a logical and rational process of data separation.  How does one maintain multi-level data separation, and at higher levels, information and knowledge separation on highly integrated networks?  Certainly the logical aspects of security can be demonstrated and local rationales for classification proven to exist in different forms.  The major difficulty arises in the maintenance and adjudication of information classification across organizational boundaries.  Assume for instance that the logical means of classification are uniform in a given system of cooperative and interactive organizations.  One organization's culture might heavily weight the classification of that information concerning the performance of information in meeting chartered objectives.  A second organization, might classify information more heavily toward the sensitivities associated with hierarchical positions.  Assume that there is a reason for these two organizations to communicate.  They have good reason to accept and use each other's data. 

 

In this scenario there are several types of classification errors that can occur which can ultimately lead to serious INFOSEC data separation compromise.  One type of classification errors would occur when information that should not have been classified at a particular level, was classified at that level or higher.  A different type classification error would occur when information that should have been classified to a particular level was not classified or was not classified to a high enough level.  In the case of the two organizational examples, there would be a high probably of both type one and type two security classification errors during  communication.  It is known that human value systems for decision making, such as make classification decisions, operate differently depending on the state of the environment.  Given that no two organizational environments can be exactly the same, classification will always present a problem for INFOSEC models as long as the classification of information is performed entirely by unaided humans.